This paper has two purposes. The first is to focus attention on the gradient nature of sound change. This characteristic of sound change, although an important one, is often overlooked. King (1969: 122), for example, states: “Phonological changes tend to affect natural classes of sounds (p, t, k, high vowels, voiced stops) because rules that affect natural classes are simpler than rules that apply only to single segments.” This perspective obscures the generalization pattern of phonological processes, for a particular process typically affects a subsection of a natural class and then may (or may not) generalize to other members of the particular class or even to other classes. The second purpose of this paper is to account for selected cases of gradient phonological change in Italian and other Romance languages on the basis of a partial theory of syllable structure preferences.

In his study of consonant epenthesis, Wetzels (1985:285-286) recognizes the existence of two distinct types: “one type involves the appearance of an epenthetic consonant inside liquid-final clusters; the second type concerns intrusive stops which originate inside sequences of consonants which do not have liquids as their second member.” Following are some characteristic examples:

In the evolution of morphological studies, morphological features and markedness have come to play increasingly important rôles. Feature notation in morphology allows the development of the notion of natural morphological class, renders explicit much of the internal structure of paradigms and permits the exploitation, suitably adapted, of many of the results obtained in phonology. Markedness considerations, particularly when linked to general theories of morphology, again highlight morphological structure, constrain the variety of permitted analyses, and indicate, at least in part, the expected direction of historical change. In the sections to follow, I will present an analysis of Old French nominal and adjectival inflection which makes key use of morphological features and the differences between the marked and unmarked values of these features. This study will be particularly concerned with constraining the morphological analysis of Old French and conversely, with using Old French data to investigate more general properties of morphological systems.